htop - convert a HMMER ASCII-formatted HMM into an equivalent
profile
htop [ -fios ] [ hmm-file | - ] [ random-model-file ] [B=#]
[C=#] [F=#] [H=#] [L=#] [M=#] [P=#] [Q=#]
htop converts a hidden Markov model (HMM) into an
equivalent PROSITE-formatted generalized profile. hmm-file is a
nucleic acid or protein HMM in HMMER ascii-format. random-model-file
contains a simple null-model in the format used by the HMM search programs
of the HMMER1 package. This parameter is only effective in conjunction with
the -b option. If omitted, a random model consisting of the average residue
probabilities of the insert states will be used.
- -f
- Emulate HMM fragment search (only effective with old HMMER1 format). When
given as input to the pftools programs pfsearch or pfscan,
the resulting profile will behave like the original HMM searched with the
HMMER1 program hmmfs. The default mode is to emulate HMM local
search implemented in the HMMER1 program hmmls.
- -i
- Force all insert extension scores to be zero.
- -o
- Assume input to be in old HMMER1 format. Pftools version 2.2 assumes by
default HMMER2 format.
- -s
- Implement semiglobal alignment mode. This option forces all external
initiation scores to be identical to the internal initiation score at the
beginning of the profile. Likewise, the external termination scores are
set to the value of the internal termination score at the end of the
profile.
- B=#
- Logarithmic base to be used for normalization of profile match scores
(only effective with old HMMER1 format). Default: B=2.0.
- C=#
- Level zero cut-off value in normalized score units. Default: C=8.5 (C=0.0
in conjunction with option -o).
- F=#
- Output score multiplier (only effective in conjunction with old HMMER1
format). On output, all profile scores are multiplied by this factor and
rounded to the nearest integers. Default: F=100. With the new HMMER2
format, which uses an integer-rounded log-odds representation of the HMM
parameters, the original values cannot be rescaled.
- H=#
- High-cost initiation/termination score (only effective in conjunction with
-s option). This score will be applied to all external and internal
initiation and termination scores at path matrix positions where
initiation or termination at low cost is not possible according to
semiglobal alignment mode. Default: H=* (low-value).
- L=#
- Logarithmic base to be used for representation of profile scores (only
effective in conjunction with old HMMER1 format). Default: 1.0233739 (1/30
bit units).
- M=#
- Maximal number of unprotected residues at the ends of the profile.
Default: M=5.
- P=#
- Percent profile-length at the ends of the profile, which will not be
included in the protected area (may be overwritten by parameter M,
see above). Default: P=0.
- Q=#
- Position-independent odds ratio assumed for unknown residues. This
parameter determines the default values for profile scores M0 and
I0. Note that the I0 scores may be over-written by the
-i option. Default: Q=0.8.
- (1)
- htop pfam_sh3.hmm > pfam_sh3.prf
Converts the HHMER2-formatted SH3 domain profile in
"pfam_sh3.hmm" into a generalized profile. After this
conversion, the following two commands:
hmmsearch --null2 pfam_sh3.hmm sh3.seq
pfsearch -fy pfam_sh3.prf sh3.seq C=5.0
should produce equivalent results in terms of raw scores and
profile-sequence alignments (E-values and normalized scores will usually
be different). Note that the program hmmsearch is from Sean
Eddy's HMMER2 package (http://hmmer.wustl.edu/).
Philipp Bucher
Philipp.Bucher@isrec.unil.ch